Many millions of lab tests are done annually across Canada. As I’ve said before, each one is an opportunity for human error leading to patient harm. Errors occur when the wrong test is ordered, or conducted on the wrong patient, or when the wrong tube is used or the specimen collection is bungled. The specimen may take too long in transport or the technologist may make a mistake during the procedure -shaking the tube too much or using the wrong chemicals, for example.

For all these reasons and more, we need champions of lab quality and safety practices. In B.C., one of those champions is Dr. Michael Noble. As I wrote in a profile on him which you can read here Noble is a giant in the Canadian microbiology field.

Today, he kicks off a new offering on this medicinematters.ca website – guest columns submitted by or solicited from top medical experts. 

Dr. Noble is in the Department of Pathology and Laboratory Medicine at the UBC Faculty of Medicine. He’s also the chair of the UBC Program Office for Laboratory Quality Management, and the Clinical Microbiology Proficiency Testing program.

 

BY DR. MICHAEL NOBLE

It’s a routine process: a person is seen by a physician who decides that a laboratory test is required to help make a decision about their medical care. A blood or other sample is collected and sent to the laboratory and in a day or so, a result comes back. The patient is told that the results indicate either the presence or absence of a specific illness.  We assume that the test result information is correct and we can be confident that the diagnosis is also correct.

But sometimes, albeit rarely, the information is wrong. And that can occur because of, for example, test errors, specimen collection, or handling.

In British Columbia, we have an electronic reporting system (Patient Safety Learning System) by which laboratories or physician offices can elect to report errors that occur. (But most do not). In the years from 2008 to 2018, over 100,000 errors have been reported. The number of voluntarily reported errors (only a minority are) per year has been rising – from about 2,000 to over 12,000 per year.  That’s a small number compared to the number of tests performed.

In most situations, errors create – at the least –  some inconvenience. But errors can also result in terrible harms. In Canada, perhaps the most infamous example of harm was the incorrect tumour marker testing in Eastern Health Newfoundland in 2005 which resulted in 126 deaths.

Laboratory errors don’t all occur in the laboratory. Many occur at the time of sample collection or transport to the lab (called pre-examination error).  Others occur because the test report was incorrect, or late, or written in a way that was misunderstood or misinterpreted (called post-examination errors).

With COVID testing, there have been problems, in part, with the science and also because the public and politicians have had a level of unblinding faith and trust in results without understanding the frailties and risks. Political pressure has forced the release of tests that we know are probably not all that good.  There are strengths and weaknesses with all testing kits and protocols.

Some errors do occur in the laboratory from insufficient quality control, or from errors during the testing process.  Sometimes, the quality control was correct and the procedure was correct but the tests that gave inappropriate results because of test inadequacies. Tests that fail to give a positive result when the sample was known to be positive are called insensitive. Tests that give a positive result when the sample was clearly known to be negative are called false positive due to non-specificity.

For example, if a test intended for COVID-19 antibodies is found to react to antibodies to other types of coronaviruses, they will test as positive, even though the patient did not have COVID-19.  Such kits were detected during a study in the United States after they had been released.  Fortunately, none have been detected in studies done in British Columbia.

Test manufacturers are required to do a lot of testing prior to release to ensure that test sensitivity and specificity are both very high, preferably each greater than 99%.

For an individual person and an individual test, having a sensitivity or specificity of 96% to 98% may seem acceptable, or even great, but when the results of 100,000 tests are combined, for example, a 4% potential error can mean 4,000 incorrect falsely negative tests and another 4,000 incorrect falsely positive tests.

Considering B.C. conducts nearly 2,000 COVID-19 2,000 samples each day and over 3,000 people have tested positive, the potential impact of faulty tests can be huge.

Many assume that if a test is ultra-scientific, results are always reliable.  Such is the case with the COVID-19 PCR test that is used to detect viral RNA in the patient’s nose.  If the sample is collected properly and tested properly, then the results can be relied upon. But the devil can be in the details; for the test to be correct, the swab has to introduced far back through the nose and to be in place for sufficient time for the virus and its RNA to be detected. Granted, it is not a pleasant test.

One day early in the outbreak, there was a physician who wanted to show how a rapid test worked on national television (FOX News – see it here). He was pleased to say that the test performed on him was negative.  But when he was tested on television, the swab was barely inside his nostril and was left there for barely a second. It’s possible he was negative but the test was likely done incorrectly and the result was probably not reliable in this scenario.

When people are rushed and procedures are not always followed, false-positive results due to contamination can occur. Contamination can occur with tests, including COVID-19, in various ways. For example, when a person collecting the sample is tending to multiple people, RNA from a previous patient can get on their gloved hands, even if they’ve washed their hands and have fresh gloves. Then they can transfer that RNA onto the swab and you may end up having two people testing positive.

Laboratory testing forms the basis of near 80% of medical decision making.  There is an obligation to ensure that they are done correctly from start to finish, to ensure that correct, accurate, timely results are the norm.

Contact: mnoble@mail.ubc.ca